Ductile deformation and mass loss in the Franciscan Subduction Complex: implications for exhumation processes in accretionary wedges

Deformation measurements from 64 sandstone samples collected in three study areas from the Eastern Belt of the Franciscan Complex are used to evaluate how the high­ pressure metamorphic interior of the Franciscan wedge was exhumed. Pressure estimates indicate 25-30 km of exhumation in this part of the Franciscan Complex. Much of the Eastern Belt has a semi-penetrative cleavage that formed by solution mass transfer (SMT) while the rocks were moving through the wedge. Individual samples have absolute princi­ pal stretches of Sx = 1 .00-1.52, Sy = 0.60-1.21 , and S2 = 0.33-0.81 . Strain magnitudes and directions are quite variable at the local scale. The deformation at the regional scale is esti­ mated by calculating tensor averages for groups of measurements. The three study areas, which are spaced over a distance of c. 500 km along the Franciscan margin, give remark­ ably similar averages, which indicates that the deformation of the Eastern Belt is consist­ ent at the regional scale. The tensor average for all data indicates a nearly vertical Z direction with Sx = 0.96, Sy = 0.92, and S2 = 0.70. Sx and Sy are near one because at the local scale, the X and Y directions vary considerably in orientation, which means that their stretch contributions are averaged out at the regional scale. This unusual strain type, con­ sisting of both plane strain and uniaxial shortening, results from the fact that shortening in Z was balanced by a pervasive mass-loss volume strain, averaging about 38% . The geometry of directed fibre overgrowths was used to measure internal rotations. These data indicate that in sandstones, SMT deformation was nearly coaxial (mean kinematic vortic­ ity number is 0.05 at the regional scale and generally <0.4 for individual samples). A simple one-dimensional steady-state model indicates that ductile thinning accounted for only c. 10% of the overall exhumation. Ductile shortening across the Franciscan wedge was very slow, at rates <8 X 10-17 s-1 ( <0.3% Ma-1). Assuming that this strain was active in an across­ strike zone <200 km wide, we estimate that horizontal ductile flow would have accounted for <0.25% of the total convergence across the Franciscan margin. We conclude that the SMT mechanism operated slowly as a background deformation process, and that the dis­ location glide mechanism was completely inactive down to depths of 25-30 km. Thus, the stability of the Franciscan wedge was probably better defined by the Coulomb wedge cri­ terion than by a viscous wedge criterion. No definitive normal faults have been found in or adjacent to the Eastern Belt. Therefore, we infer that wedge taper was mainly controlled by deep accretion and erosion of an emergent forearc high. A fundamental problem in tectonics is the cause of deep exhumation of high-pressure metamor­ phic rocks commonly found in the interior of many convergent wedges. In recent years, exten­ sional faulting has received considerable atten­ tion because it provides an elegant and possibly widely applicable mechanism for unroofing a wedge (e.g. Platt 1986, 1993; Dewey 1988). In addition to normal faulting, erosion (e.g. England & Richardson 1977; Rubie 1984; Brandon et al. 1998) and syn-convergent ductile flow (e.g. Selverstone 1985; Wallis 1992; Wallis et al. 1993; Feehan & Brandon 1999) can also con­ tribute to exhumation of a convergent wedge. Three different settings of deep exhumation have been recognized. One is in the internal zones of continent-continent collisions where exhumed high-pressure metamorphic rocks, including both high-temperature and low­ temperature varieties, are found. This setting contains some of the deepest exhumed rocks, coesiteand diamond-bearing rocks coming from depths >100 km (e.g. Chopin 1984; Coleman & Wang 1995). A second setting is found in ancient subduc­ tion complexes and is characterized by exotic blocks of blueschist and eclogite immersed in a matrix of highly deformed mudstone or RING, U. & BRANDON, M. T. 1999. Ductile deformation and mass loss in the Franciscan Subduction Complex: Implications for exhumation processes in accretionary wedges. In: RING, U., BRANDON, M. T., LISTER, G. S. & WILLETT, S. D. (eds) Exhumation Processes: Normal Faulting, Ductile Flow and Erosion. Geological Society, London, Special Publications, 154, 55-86. 56 U. RING & M. T. BRANDON serpentinite. These blocks typically come from depths of 30-60 km. A classic example of this setting is the knocker terrane of the Central Belt of the Franciscan Complex. The process of exhumation remains poorly understood. Platt (1986) favoured normal faulting, whereas Cloos & Shreve (1988) proposed upward transport by deep-seated channelized flow at the base of the wedge. We focus here on a third setting, where large coherent tracts of low-temperature-high­ pressure metamorphic rocks have been ex­ humed in subduction-related accretionary wedges. The metamorphic rocks involved are typically of prehnite-pumpellyite, lawsonite­ albite, pumpellyite-actinolite, blueschist, or more rarely greenschist facies. The coherence of these terranes is indicated by the common preservation of metamorphic isograds within the terranes. The depth of exhumation is usually no greater than about 35 km, which is consistent with the maximum thickness of modern subduc­ tion-related wedges. Normal faulting is often cited as a likely exhumation process in this setting. The island of Crete, which marks the modern forearc high of the Hellenic convergent margin, provides an example of syn-convergent normal faulting (Fassoulas et al. 1994; Thomson et al. this volume). Normal faulting there seems to be controlled by roll-back of the Hellenic sub­ duction zone. Platt (1986, 1987), Jayko et al. (1987), and Harms et al. (1992) have argued that coherent low-temperature-high-pressure meta­ morphic terranes in the Eastern Belt of the Franciscan Complex were unroofed by syn­ convergent normal faulting. In contrast, erosion appears to be the primary exhumation process for the modern forearc high at the Cascadia margin (Brandon et al. 1998). In this paper, we report new deformation measurements from the coherent metamorphic terranes of the Eastern Belt of the Franciscan Complex and use this information to resolve the deformation field that existed within the Fran­ ciscan wedge. The high-pressure metamorphic rocks of the Eastern Belt show extensive evi­ dence of solution mass transfer (SMT) defor­ mation. The clockwise P-T loop for these rocks indicates a general displacement path involving subduction, then accretion at the base of the wedge, upward flow within the wedge, followed by exhumation and exposure at the Earth's surface. We maintain that these deeply exhumed rocks provide a path-integrated record of the deformation-rate field in the wedge. We utilize some new methods, which provide a full determination of the absolute finite deformation produced by the SMT mechanism, including volume strain and internal rotation. These methods are the projected dimension strain (PDS) method, the mode method and the semi­ deformable antitaxial (SDA) fibre method. The contribution of vertical ductile thinning to exhumation is estimated using a simple one­ dimensional steady-state model by Feehan & Brandon (1999). The data reported herein, com­ bined with our work on the kinematic evolution of the Coast Range fault zone (Ring & Brandon 1994, 1997), allow us to approximate the relative contributions of different processes to the total exhumation of the Eastern Belt metamorphic rocks.